Superficially porous particles with carbon core and nanodiamond polymer shell for protein separations. Szabolcs FEKETE, Davy GUILLARME

Transcription

1 Superficially porous particles with carbon core and nanodiamond polymer shell for protein separations Szabolcs FEKETE, Davy GUILLARME

2 Current trends in column technology Core-shell small particles ( µm) N max ~ plates/meter Fully porous small particles ( µm) N max ~ plates/meter Silica monolith (2 nd generation) N max ~ plates/meter Organic polymer monoliths N max ~ plates/meter Close to columns are commercially available for HPLC! very fast very efficient very high pressure Parallel segmented flow PLOT columns Silica colloidal crystals Pillar arrays Ordered structures

3 New carbon-diamond based superficially porous material FLARE columns - Unique selectivity relative to silica - ph stability: 1 < ph < 13 - Thermally stable up to 100 ºC diamond-analytics.com

4 Other possible applications? For large biomolecules?! 1) It possesses an average pore size of 250 Å 2) No residual silanols 3) Retention mechanism and selectivity are different compared to silica based materials (hydrophobic + anion-exchange) 4) It is stable at high temperature (recovery, effective molecular diffusion) 5) Diamond and carbon have advantageous heat properties (important for large solutes retention control) less frictional heating no negative impact on retention and band broadening 6) It offers the benefits of core shell (pellicular) particle technology

5 Column efficiency, plate height equations h = f (ν ) The old empirical formula (1956): b h = a + + ν cν More sophisticated recent models (Knox, Golay, Giddings, Horvath, Huber, Myabe, Guiochon ): h = aν b h = aν c fν + cpν + ν b + + c f ν + c p ν + ν h heat ( 1/ 2λ ) + ( 1/ 2ω ν ) heat 1 a eddy dispersion b - longitudinal diffusion c mass transfer resistance c f - film or external c p - transparticle or internal heat eddy dispersion coefficient related to a flow exchange mechanism generated by heat friction in the column eddy dispersion coefficient related to a diffusion exchange mechanism (Aris) generated by heat friction in the column

6 Why superficially porous particles? Core-shell advantages: 1) More advantageous C p term (for large molecules) (Horváth in the late 60 s) 2) lower B term (for small molecules) ρ = R i /R e 3) lower A term... Compromises: - Decrease in retention - Lower loading capacity Optimal structure: Compromise, 0.6 < ρ < 0.9 (porous volume fraction between 80 and 20 %)

7 The impact of particle structure volume fraction expected gain in efficiency loadibility phase ratio retention 3.6 µm 1 µm ρ = 0.94 FLARE FLARE volume fraction ~16% pellicular S. Fekete, D. Guillarme, M. Dong, LC-GC NA, 32 (2014) 2.

8 Theoretical h ν plots Theoretical h-ν curves of fully porous and core-shell packing (ρ=0, ρ=0.63, ρ=0.87 and ρ=0.73). Hypothetical mobile phase composition: 50 % acetonitrile 50 % water, and analyte molecular weight: 1000 g/mol. No extra-column band broadening is assumed.

9 The effect of molecular diffusion on plate height reduced plate height B-term region For proteins, the C-term is the most important contribution steroid insulin 10 kda protein reduced linear velocity

10 Gradient elution, peak capacity Peak capacity describes the column performance in gradient elution mode gradient time P tg = w 50% peak width at half height Corresponds to a resolution of 1 between consecutive peaks In analogy with isocratic plate height (H) equations, when combining isocratic plate height and gradient peak capacity equations the next formula can be written: P = 1 4 A+ ((2γD M t 0 L ) / L) + C(1/ D S )( L / t 0 ) 1 k ln G + 1 k 0,max 0,min retention window A B C gradient slope the effect of molecular diffusivity and flow rate (that can be expressed with t 0 ) on the peak capacity can directly be highlighted.

11 Peak capacity with model proteins on FLARE 100 x 2.1 mm FLARE column, mobile phase A: 0.2% TFA, mobile phase B: 0.2% TFA in ACN, gradient: %B Effect of temperature Cytochrome c TRFE BSA F = 0.2 ml/min, T 1 = 30 ºC, T 2 = 50 ºC Effect of flow-rate Cytochrome c TRFE BSA T = 30 ºC, F 1 = 0.2 ml/min, F 2 = 0.4 ml/min B. Bobály, D. Guillarme, S. Fekete, J. Pharm. Biomed. Anal. 104 (2015) 130.

12 Impact of TFA% FLARE column increase in retention improvement in peak capacity F = 0.2 ml/min, T = 30 ºC AERIS WP silica-based Cytochrome c TRFE No difference between 0.1 and 0.3% TFA F = 0.2 ml/min, T = 30 ºC BSA - decrease in overall polarity (increase in hydrophobicity) of the proteins as the TFA concentration increased - decrease in polarity of the stationary phase through ion pairing effects of TFA % TFA is required B. Bobály, D. Guillarme, S. Fekete, J. Pharm. Biomed. Anal. 104 (2015) 130.

13 Temperature, additive, flow-rate Differences compared to silica-based conventional columns No need to work at high temperature (avoiding on-column protein degradation) Fast gradients can be applied (without loosing peak capacity) Higher amount of TFA is required ( %) Useful for fast gradient protein separations performed at low temperature!

14 Real life protein separations Why using RP-HPLC? Assay (protein content) Protein impurities Degradants of proteins (oxidation, reduction, deamidation) Small Aggregates (dimer) Heterogeneity (charge, isoforms) Heavy and light chain variants (mab) Fab and Fc variants, fragments (mab) Investigation of disulphide bridges Excipients (polysorbate-20, -80) Amino Acid Analysis (AAA) Peptide maps Closely related proteins have to be separated! Method development R = N α 1 k 4 α k + 1 Ø Very efficient recent columns Ø Optimization of column length Ø Increasing temperature Ø Organic modifier Ø ph Ø Temperature

15 Separation of interferon variants 19.2 kda Column: FLARE 100 x 2.1 mm C18 M, Mobile phase A : 0.3 % TFA, B : 0.3% TFA in ACN, gradient: %B in 7 min, flow rate: 0.3 ml/min, temperature: 30 C. Detection: fluorescence (λex: 280 nm, λem: 360 nm, 20 Hz sampling rate). B. Bobály, D. Guillarme, S. Fekete, J. Pharm. Biomed. Anal. 104 (2015) 130.

16 Separation of filgrastim vaiants 18.8 kda Column: FLARE 100 x 2.1 mm , Mobile phase A : 0.5 % TFA, B : 0.5% TFA in ACN, gradient: %B in 4 min, flow rate: 0.2 ml/min, temperature: 60 C. Detection: UV (280 nm, 20 Hz sampling rate).

17 Filgrastim pharmacopeia method (related proteins) Reduction of the analysis time by a factor of 6 And solvent consumption by a factor of 25 United States Pharmacopeia (USP)

18 Peak capacity (filgrastim) Impact of gradient time (steepness) at 60 ºC Impact of temperature at 6.25 %B/min gradient steepness (4 min long gradient) FLARE 250A FLARE 180A FLARE 250A FLARE 180A Mobile phase A : 0.5 % TFA, B : 0.5% TFA in ACN, gradient: %B (on FLARE) and %B (on AERIS), flow rate: 0.2 ml/min. Columns: 100 x 2.1 mm

19 Analysis of mabs mabs represent an important class of therapeutic proteins with a wide range of clinical indications Chemical and ezymatic modifications during manufacture, formulation, storage Intrinsic microheterogeneity Common modifications: N-glycosylation Methionine oxidation Fragmentation Lysine truncation Deamidation Fab amino termini variants (glutamine pyroglutamate) Their detailed characterization requires several orthogonal methods/tools Among them RPLC is an important tool.

20 Strategies, approaches in LC for mab analysis Ø 1 Complete proteolytic digestion of a mab (peptide mapping, bottom-up approach) Ø 2 Analysis of fragments/domains: HC, LC, Fab, Fc, F(ab ) 2, sfc, Fd ( middle down ) Ø 3 Analysis of intact mabs ( top down ) Reversed phase (RP-HPLC): intact mab (heterogeneity, similarity) HC-LC (oxidation, variability) Fab-Fc, sfc-fd (variability, post-translational modifications, such as N-terminal cyclization, oxidation, deamidation, and C-terminal processed lysine residues) Peptide mapping (primary structure) Large hydrodynamic radii ( 35Å) Possible issues Huge number of charges ( 100) Low diffusion coefficient Peak tailing Peak broadening AdsorpDon

21 Recovery of intact mab (rituximab IgG1) 147 kda Column: BEH300 C4 (150 mm x 2.1 mm), temperature: from 40 ºC up to 80 ºC, injected volume: 0.5 µl, detection: 280 nm. Mobile phase A: 0.1% TFA in water, mobile phase B: 0.1% TFA in acetonitrile. Gradient: from 30% to 37% B in 6 min, flow rate: 0.3 ml/min. S. Fekete, S. Rudaz, J.L. Veuthey, D. Guillarme, J. Sep. Sci. 35 (2012) 3113.

22 Intact and reduced mab (bevacizumab IgG1) LC (~25 kda) + HC (~50 kda) Variants of the intact mab can be separated. Column: FLARE 100 x 2.1 mm C18 M, Mobile phase A : 0.2 % TFA, B : 0.2% TFA in ACN, gradient: %B in 4 min, flow rate: 0.3 ml/min, temperature: 90 C. Detection: UV (280 nm, 20 Hz sampling rate).

23 Heavy chain variability (IgG2 panitumumab) HC LC HC variants Column: FLARE 100 x 2.1 mm C18, Mobile phase A : 0.5 % TFA, B : 0.5% TFA in ACN, gradient: %B in 10 min, flow rate: 0.15 ml/min, temperature: 90 C. Detection: FL ( nm).

24 Trastuzumab (IgG1) papain digested sample Fab variants Fab (~50 kda) Fc (~50 kda) Fc + Column: FLARE 100 x 2.1 mm , Mobile phase A : 0.5 % TFA, B : 0.5% TFA in ACN, gradient: %B in 4 min, flow rate: 0.2 ml/min, temperature: 90 C. Detection: UV (280 nm, 20 Hz sampling rate).

25 Papain digested + reduced IgG1 (rituximab) Fd sfc (~25 kda) + + HC Fd (Fab portion of HC) LC (~25 kda) LC sfc papain digested + reduced reduced intact Column: FLARE 100 x 2.1 mm C18, Mobile phase A : 0.5 % TFA, B : 0.5% TFA in ACN, gradient: %B in 10 min, flow rate: 0.15 ml/min, temperature: 90 C. Detection: FL ( nm).

26 ADC, digested sample Brentuximab-Vedotin was digested with endoproteinase Lys-C Column: FLARE 100 x 2.1 mm , Mobile phase A : 0.5 % TFA, B : 0.5% TFA in ACN, gradient: 0-50 %B in 12 min, flow rate: 0.2 ml/min, temperature: 60 C. Detection: UV (280 nm, 20 Hz sampling rate). Column coupling

27 Conclusion This carbon-based nanodiamond-polymer shell material is interesting for the separation of large molecules With most proteins, there is no need to work at high temperature When high temperature is required (e.g. mabs), it is again useful since it can be operated at up to 100 ºC The achievable peak capacity is comparable to the best silica-based wide-pore materials The retention is somewhat lower compared to silica-based fully porous and core-shell materials but can easily be adjusted by decreasing the solvent-strength or by the higher amount of TFA

28 Acknowledgments Prof. Jean-Luc Veuthey Prof. Serge Rudaz Balázs Bobály Alain Beck Imre Molnár Andrew Dadson Thank you for your attention!